Articulo HFM

Published July 1, 1964
A MACROMOLECULAR R E P E A T I N G U N I T OF M I T O C H O N D R I A L STRUCTURE AND F U N C T I O N Correlated Electron Microscopic and Biochemical Studies of Isolated Mitochondria and Submitochondrial Particles of Beef Heart Muscle
H. F E R N A N D E Z - M O R A N , M.D., T. O D A , M.D., P. V. B L A I R , Ph.D., and D. E. G R E E N , Ph.D. From the Mixter Laboratories for Electron Microscopy, Neurosurgical Service, Massachusetts General Hospital, Boston, and The Department of Biophysics, University of Chicago; and the Institute for Enzyme Research, University of Wisconsin, Madison. Dr. Fem~indez-Mor~n's present address is Department of Biophysics, University of Chicago. Dr. Oda's present address is Department of Pathology, Okayama University Medical School, Olmyama, Japan
Downloaded from jcb.rupress.org on February 4, 2011
ABSTRACT A repeating particle associated with the cristae and the inner membrane of the external envelope has been recognized and characterized in beef heart mitochondria by correlated electron microscopic and biochemical studies. Many thousands (ca. 104 to 105) of these particles, disposed in regular arrays, are present in a single mitochondrion. The repeating particle, called the elementary particle (EP), consists of three parts: (1) a spherical or polyhedral head piece (80 to 100 A in diameter); (2) a cylindrical stalk (about 50 A long and 30 to 40 A wide); and (3) a base piece (40 X 110 A). The base pieces of the elementary particles form an integral part of the outer dense layers of the cristae. The elementary particles can be seen in electron micrographs of mitochondria in situ, of isolated mitochondria, and of submitochondrial particles with a complete electron transfer chain. Negative staining with phosphotungstate is only one of several techniques that can be used for reproducible demonstration of the repeating particles and underlying subunit organization of mitochondrial membranes. A particulate unit containing a complete electron transfer chain can be isolated from beef heart mitochondria. The isolated unit approximates in size that of the elementary particle in situ. The molecular weight of the particle in situ is calculated to be 1.3 X 10~. Evidence is presented for identifying the isolated unit with the elementary particle visualized in situ. The elementary particle of the mitochondrion is believed to be a prototype of a class of functional particles or macromolecular assemblies of similar size found in association with membranes generally.
INTRODUCTION
The pioneering electron microscope studies of Palade (53, 54) and Sj6strand (67, 68) established more than a decade ago the distinctive fine struc-
ture of the mitochondrion as essentially that of an elongated subcellular body bounded by a membrane consisting of two layers. The external layer
63
is considered to be the limiting membrane, while regular infoldings of the inner membrane form numerous internal ridges termed "cristae" by Palade. During the same period the mitochondrion was identified as the locale of citric cycle oxidations, electron transfer, and oxidative phosphorylation (49, 65). The concept of the mitochondrion as an organized system in which there was a precise arrangement of the many enzymes involved in the sequential reactions underlying mitochondrial function (32, 36, 50) became firmly established in biochemical thinking. Large scale isolation of stable mitochondria (33, 51) under conditions which do not impair their main enzymic activities has made these highly organized, membranous organelles available for chemical analysis. Equally important, there are multiple, exact criteria by which the functional integrity of the mitochondrion can be evaluated. Of all the specialized lamellar systems, mitochondria appeared, therefore, to be particularly well suited for a study of correlations between biochemical function and ultrastructure (2325, 37). Respiratory energy transformations have been shown to take place in the organized membrane structures of the mitochondrion. The mitochondrial membranes contain the different multienzyme complexes involved in these transformations; these complexes are arranged in highly ordered arrays (33, 37, 40, 51). Although the first phase of the fine structural analysis of the mitoehondrion (53, 54, 67) established the base line for a correlation of structure and function (4, 14, 65, 79, 80), the shortcomings of the preparative techniques and of the resolution of the electron microscope limited the extent of meaningful correlation. With mitochondria, as with other lamellar systems (18, 68, 78), it is now apparent that the submicroscopic patterns of sectioned specimens fixed with osmium tetroxide reveal merely a general structural framework consisting essentially of lipoprotein. Thus, in the mitochondrial membranes only certain stereotyped, uniform features of the "osmium-stained" smooth membranes are seen, with no indications of the specific enzymic complexes and other constituents. The mitochondrion is a lamellar system in which the biochemical and enzymic properties are better characterized than is the ultrastructure. Recent improvements in preparative techniques, particularly that of negative staining
applied to the study of virus ultrastructure (7, 8), have made possible the examination of biological systems in far greater detail than could heretofore be achieved in sectioned material. Fortunately, mitochondria (both those seen in situ and after isolation), by virtue of their ultrathin membranes, are particularly suitable for application of many of these new techniques which do not necessarily require sectioning (16, 23). Three years ago, we applied these techniques, including negative staining and improved low temperature methods which yield better morphological and histochemical preservation in sectioned material (20-23), to the study of mitochondria. These techniques, combined with improved high-resolution electron microscopy, revealed for the first time the presence of a characteristic polyhedral or round structural unit, 80 to 100 A in diameter, as a basic component of mitochondrial membranes (23, 25, 26). This finding casts a new light on our views of the organization of the mitochondrion. Recognition of this repeating particulate subunit, which we designated "elementary particle" (EP), was the starting point for a program of biochemical isolation. In essence, the experimental approach of our studies correlating structure with function involved fragmentation of mitochondria into subunits. The mitochondrion was, in effect, treated as a chemical entity capable of stepwise disassembly into its component parts (and reassembly into the original unit) (40). This approach, from a biochemical point of view, has already led to the discovery of new electron transfer components such as coenzyme Q, and of proteins containing non-heme iron and copper. The study of electron transfer in intact cells and in particles containing the respiratory chain (as exemplified by the studies of Chance and his collaborators (9, 10) has been of great value in promoting understanding of the over-all kinetics and the sequence of oxidoreduction of the spectroscopically visible components of the respiratory chain. A third approach involved reconstitution of a fully functional electron transfer particle from its purified segments and from the highly purified and independent component parts (37, 44, 45). Studies on the stepwise degradation of the mitochondrion into particles of decreasing complexity have led to the recognition of three basic components of mitochondrial organization: (1) the particulate electron transfer chain; (2) the
64
THE JOURNALOF CELL BIOLOGY" VOLUME~ , 1964
Insoluble network composed of structural protein and lipid; and (3) the "solubilizable" dehydrogenase'complexes. This was the biochemical picture at the time when the collaborative effort of our two laboratories began. Sonic irradiation dissociates mitochondria into a particulate structured fraction and a soluble fraction; this resolution makes it clear that there cannot be a single mitochondrial subunit which is all-embracing, but rather several subunits depending upon the function under consideration. We are particularly concerned here with the subunit that contains the complete electron transfer chain--the apparatus for the transfer of electrons from succinate and DPNH to molecular oxygen. We have worked mainly with mitochondria prepared from beef heart muscle, because they are relatively stable and can be isolated on a large scale; an additional advantage of heart mitochondria over liver mitochondria is their higher density of cristae per unit volume with a correspondingly greater oxidative rate. By careful comminution of beef heart mitochondria in a sucrose medium, well defined fractions were prepared, one of which (ETP) contains particles with an essentially intact electron transfer chain, divested of the primary dehydrogenating enzyme complexes. The capacity for both electron transfer and oxidative phosphorylation was retained in an analogous particle (ETPI~). Electron microscopic studied-of these submitochondrial fractions by standard osmium fixation and thin sectioning techniques (14, 33, 79, 80) indicated that the double-membraned structure was preserved in specimens of ETPm whereas only a single-membraned structure was observed in ETP. More extensive correlation of fine structure with function, however, could not be achieved in these early investigations. Concentrating on the electron transfer particle (ETP) as the primary unit of mitochondrial function, Green and his colleagues (33, 37, 40) succeeded in isolating and characterizing eleven oxidation-reduction components of the electron transfer chain. At least nine proteins with oxidation-reduction groups participate in the terminal electron transfer process. Approximately 30 per cent of the total dry weight of the mitochondrion and of the ETP is accounted for by lipids, the bulk of which (95 per cent) is phospholipid (29 b). Green and Oda (38) tentatively assumed, from preliminary data, that the structural unit of
ETP was a cylindrical particle about 445 A long and 100 A in diameter, with a particle weight of the order of 4.3 million. Three developments have pointed to a downward revision of the estimated value proposed (38) as the molecular weight of the repeating unit of the electron transfer particle: (1) the demonstration by Fernfindez-Morfin (23, 25) that the membrane structure of isolated mitochondria and subfractions appeared to be built up of paired arrays of the 80 to 100 A particles, noted above, separated by a "middle" or mesolayer of variable width; (2) the isolation by Green et al. (12, 13, 39) of a structural protein, devoid of oxidation-reduction groups, accounting for 60 to 70 per cent of the total mitochondrial protein; (3) the resolution of the electron transfer chain into four complexes functionally independent (44, 45), each of high purity, and the reconstruction of the original particle with essentially undiminished activity by recombination of the four complexes carried out by Hatefi and his colleagues (30, 44, 45). The postulated molecular weight of 4.3 106 for the electron transfer chain was hardly compatible with the dimensions of the 80 to It0 A repeating unit seen by Fernfindez-Morfin. ~Ihe observation that a large proportion of the total pro~ein of the mitochondrion was not concerned with the electron transfer process per se suggested the possibility that the electron transfer particle as isolated could be resolved into a fraction containing structural protein and a fraction containing the electron transfer chain. The latter unit would have a calculated molecular weight of 1 to 2 106; this value would bring the size of the unit better into line with~he dimensions of the repeating particle seen by the electron microscope. The particle weight of the reconstructed electron transfer chain (from the four purified complexes) was estimated to be 1.4 X 106 on a protein basis, calculated from the known particle weights of each of the four constituent complexes (established from composition and ultracentrifuge data). These developments made the conclusion inescapable that the n~olecnlar weight of the unit of electron trans-fer had to be no more than 1 to 2 X 106 and led us to a program directed toward isolation of a unit of electron transfer of molecular weight less than that of ETP. Indeed, such a particle has been isolated (5). It contains the complete electron transfer chain with all the components in the same molecular ratios as those found in the mitochon-
FERN~.NDEz-MoR~.N ET AL. Macromolecular Repeating Unit of Mitochondrial Structure
65
drial electron transfer chain. Increase in activity a n d increase in concentration of the oxidationreduction proteins per unit weight run parallel throughout all but the last stage in the purification. T h e molecular weight of the particle at the highest purity level attained is 1.4 X l06 on a protein basis (this would correspond to a molecular weight of 2 X l0 s for the lipoprotein complex). T h e particle reconstructed from the four isolated complexes has a molecular weight that corresponds closely to that of the integrated particle isolated directly from the mitochondrion. I n consequence of these developments, we have elected to designate the particle obtained both by isolation and by reconstruction as the elementary particle, and we have postulated that the isolated or reconstructed elementary particle corresponds to the repeating particle with a head piece of 80 to 100 A, a stalk, and a base piece, seen in electron micrographs. After the initial phase was completed, resulting in a definition of the basic structural parameters of the new repeating particle and the establishm e n t of its relationship to the mitochondrial m e m branes, it r e m a i n e d to be established that the visualized particle is a bona fide constituent of the m e m b r a n e system. Mitochondria from a wide variety of sources were examined and the preparative procedures for the elecl-mn microscopic examination were carefully checked for the possibility of artifact formation. In addition, a complete set of particles was examined ranging in size from the mitochondrion to the individual complexes; a n internal control of artifact formation thus became available since only particles with a complete electron transfer chain could be expected to constitute the repeating unit. T h e present communication will deal specifically with: (1) the electron microscopic evidence of the existence of a repeating particle in whole mitochondria and in those submitochondrial particles that have intact electron transfer chains; (2) the isolation and properties of the particles; (3) the experimental basis for identifying the isolated or reconstituted particle with the repeating particle visualized by the electron microscope. Preliminary accounts of facets of the present study have been reported elsewhere (23-27, 34). MATERIAL~ AND METHODS
MATERIALS : i n the course of extensive correlative studies, mltochondria from many sources were
examined either in situ or after isolation from heart muscle, retina, liver, pancreas, and brain of the rat, mouse, beef, chicken, and guinea pig; but the principal effort was directed to the study of isolated beef heart rnitochondria. I S O L A T I O N OF M I T O C H O N D R I A : Large scale isolation of mitochondria from beef heart muscle was carried out under conditions that minimized contamination by other cell particulates and by myosin (5, 11). Suspensions in 0.5 M sucrose (kept at 0 to 5 ) conformed to the most rigorous standards for preservation of mitochondrial structure and activity. Two alternative preparative procedures were used. In the first procedure ground beef heart muscle is homogenized in 0.5 M sucrose at pH 7.0 and the mitochondrial fraction is separated from other structures and compounds by differential centrifugation. In turn, the rnitochondrial fraction is refractionated; a heavy fraction containing only intact whole mitochondria with well preserved cristae is collected. For ease of presentation the details of this preparative procedure are summarized in Table I. In the second procedure ground beef heart muscle was suspended in 0,66 M sucrose and treated with Nagarse proteinase according to the method of Hagihara (43). In the klagihara method the use o f the high speed blendor is eliminated. The proteolytic enzyme attacks the myofibrils and connective tissue and thus releases mitochondria from the muscle mass. We have found no recognizable difference in the purity or integrity of the mitochondria prepared by the two procedures outlined above. Both types of preparations were used for electron microscopy of intact mitochondria, Freezing of the suspensions was avoided and generally no more than 8 to 24 hours intervened between the preparation of the mitochondria and examination by electron microscopy, ELEMENTARY PARTICLES: The detailed procedure for the isolation of elementary particles has been described in the companion article by Blair et al. (5). A summary of the preparative procedure is given in Table II. It should be pointed out that elementary particles isolated from the standard preparation of mitochondria in 0.25 M sucrose are equally as satisfactory as those prepared from mitochondria that had been purified more rigorously for good electron microscopy. The mitochondrial suspensions which served as the starting point for the isolation of EP had "theoretical" P / O ratios (the value was two for succinate and three for pyruvate plus malate). Furthermore, they showed all the characteristics of intact mitochondria: nonreduction of external cytochrome c or DPN +, nonoxidation of external DPNH, non-oxidation of citrate or isocitrate, low rate of oxidation for succinate and negligible ATPase activity. Essentially the isolation of the elementary particle
66
THE JOURNAL OF CELL BIOLOOY - VOLUME ~ ,
1964
involves a quantitative separation of the mitochondrial subunits; this separation is achieved by ammonium sulfate fractionation of mitochondria that have been frozen in presence of 0.3 per cent KCI washed with 0.9 per cent KC1, and finally "solubilized" with a mixture of cholate and deoxycholate. The insoluble structural protein fraction sediments at low concentrations of ammonium sulfate; the floating fraction containing the elementary particle separates at about 0.5 saturation; the soluble fraction (not precipitated by 0.5 saturation) contains
some of the primary dehydrogenase complexes and cytochrome c. When the procedure outlined in Table II is repeated, the elementary particle thus obtained is stripped further of contaminating structural protein and is correspondingly enriched with respect to the oxidation-reduction components. Thus, the cytochrome a content of the one-cycle material is about 2.8 re#moles per mg of protein and of the two-cycle material as high as 4.2 re#moles per rag of protein. For ultracentrifugal and x-ray scattering analyses,
TABLE I
Method/or Large Scale Preparation of Mitochondria in 0.5 M Sucrose

1. Suspend 3200 gm of ground beef heart muscle in 9.6 1. of 0.5 M sucrose containing 0.02 M K2HPO, (final volume ca. 13 1.). 2. Adjust pH to 7.0 with N K O H and filter the suspension through cheese cloth (double layer). The filtered residue is resuspended in 9.6 1. of 0.5 M sucrose containing 0.02 M phosphate buffer of pH 7.0. 3. Homogenize in macro Waring blendor for 1 min. with the rheostat setting at 60. Neutralize the suspension with N K O H to pH 7.0. 4. Two fractions are removed after centrifugation at 0-5 : R l - - a f t e r centrifugation for 20 rain. at 1600 g (13-1iter refrigerated centrifuge) ; R2--after centrifugation at 50,000 g in the Sharpies continuous centrifuge. 5. R~ is resuspended in 1 1. of a mixture 0.5 M in sucrose and 0.01 g in phosphate (pH 7.0); the homogenized suspension is centrifuged at 15,000 g for 20 rain. The residue consists of a well packed as well as a loosely packed layer. The two layers are separated, suspended in about 500 ml of the sucrose-phosphate medium, and the homogenized suspensions are recentrifuged at 15,000 g for 20 min. The suspension prepared from the well packed residue layer on centrifugation yields a predominantly well packed layer topped by a small layer of loosely packed material. Only the well packed material is resuspended in 0.5 ~ sucrose and this suspension constitutes the final mitochondrial preparation.
T A B L E II
Summary of the Procedure/or Isolation of Elementary Particles from Beef Heart Mitochondria (Blair et al., 5)
1. Standard large scale isolation of beef heart mitochondria by the method of Crane et al. (9). Suspension medium~0.25 ~ sucrose. 2. Freezing of mitochondria in a mixture containing 0.3% KC1 and 0.17 M sucrose and then washing the thawed rnitochondria in 0.9% KC1. Residue resuspended in 0.66 M sucrose to a final sucrose concentration of 0.3 M. 3. Mitochondria exposed to deoxycholate (0.3 nag per mg protein) and cholate (0.3 rag/rag protein) at p H 8.0 (0.02 M Tris buffer) and fractionated with saturated ammonium sulfate. The first fraction at 33 per cent saturation was discarded; the second at 50 per cent saturation (floating pellet) was retained. 4. Floating pellet diluted to 0,5 mg of protein per ml in 0.25 M sucrose and sedimented at 30,000 RPM (4 hours). Pellet resuspended in buffered 0.25 M sucrose; the suspension was clarified by sonic irradiation. Insoluble material was removed by centrifugation. The clear supernatant solution contains a fine suspension of the elementary particles.
FERN.~NDEZ-MOR,~N ST AL. Macromolecular Repeating Unit of Mitochondrial Structure
67
the suspension of elementary particles (two cycles of treatment; 20 mg of protein per ml) was "solubilized" with 1.5 X 10-3 M sodium phosphotungstate in presence of 0.2 per cent a-tocopherol (to minimize lipid peroxidation). The particles were separated into a tightly packed sediment which was discarded, and a translucent liquid precipitate which was used for physical measurements as well as for electron microscopy. COMPLEXES OF THE ELECTRON TRANSFER CHAIN: The four complexes that make up the electron transfer chain were prepared by the following procedures: succinic-coenzyme Q reductase by the method of Ziegler and Doeg (81), or of Tisdale et al. (75 a); DPNH-coenzyme Q rcductase by the method of Hatefi et al. (47); QH2--cytochrome reductase by the method of Rieske, Zaugg, and Wharton (61); and cytoehrome oxidase by the method of Griffiths and Wharton (42) or that of Wharton and Tzagoloff (77). These four complexes interact to form a reconstituted elementary particle under the conditions specified by Hatcfi et al. (44, 45) and by Fowler and Richardson (30). ELECTRON TRANSFER PARTICLE: The clectren transfer particle with phosphorylating properties (ETPn) was prepared from heavy beef heart mitochondria by the method of Smith and Hansen (69). METHODS OF E L E C T R O N
The principal preparation procedures used can be classified as follows: 1. Modified negative staining techniques requiring minimal preparative manipulations were mainly used. These techniques do not involve sectioning, but only extension or surface spreading of whole mitochondria and mitochondrial membranes. 2. Thin sectioning of mitochondria by standard and modified techniques for fixation and embedding. 3. Cryofixation and related low temperature preparation techniques. Most of the results described in this report were obtained with methods of group 1, which will be described in greater detail.
1. Preparation
of
Mitochondria by Modified
Negative Staining Techniques

(a) Isolated mitochondria suspended in microdroplets of a sucrose solution containing 0.1 to 1 per cent potassium phosphotungstatc (pH 7.2) are sandwiched between carbon-coated plastic films, or impermeable single-crystal graphite, or mica lamellae in vacuum-fight microchambers of special design (20-25). When used in combination with low-intensity microbeam illumination and controlled specimen cooling (0 to --130C), this technique for electron microscopy of wet or partly hydrated biological systems yielded the first useful pictures of the regular mitochondrial membrane particles (23). Cytochrome c can he used instead of phosphotungstate to enhance contrast through "negative staining" by embedding flattened mitochondrial membrane
MICROSCOPY A variety of complementary methods was applied in the course of the correlated studies of mitochondria.
Index to Electron Micrographs

NOTEWORTHY ITEM*
FIGUR]~ NUMBER 1, 2, 3 4, 5, 6, 7, 8, 9, 15, 27 14 10, 11, 12, 13 5, 7, 8, 9, 10, 12, 13 17, 18, 19, 20, ~8 21, 22, 29 23, 24, 25, 26
Subunit structure in sections cristae Arrays of EP in isolated mitochondria Arrays of EP in ETP (Electron transfer particle) Structure of head piece EP Structure of stalk and base piece EP Isolated elementary particle (EP) Reconstituted elementary particle Complexes of electron transfer chain * See arrows.
FmvaEs 1 and 2 a Thin sections of isolated beef hcart mitochondria, "Nagarse" preparations (0.66 sucrose); osmium fixation, low temperature embedding in methacrylate, stained with Pb(OH)2. Note regularity of cristae and preservation of fine structure. Fig. 1, X 180,000; Fig. 2 a, X 180,000. FmUItE 2 b Enlargement of Fig. ~ a demonstrating particulate subunit structure in dense outer layers of each crista. X 650,000. Fiorv-nE 3 Enlarged segment of loop of a crista in a specimen similar to that shown in Fig. 2 a. Clearly visible array of globular subunits about 70 to 80 A in diameter. X 500,000.
68
THB JOURNAL OF CELL BIOLOGY VOLUME 22, 1964
FERN.~NDEz-MoR~.N EL AL.
Macromolecular Repeating Unit of Mitochondrial Structure
69
in the ultrathin liquid layers (ca. 500 to 1000 A) which are enclosed between the microchamber film "windows." Since this technique essentially yields negative staining of wet specimens without most of the usual drying artifacts, it is proving to be indispensable for studies of labile membrane structures and lipoprotein systems under conditions approaching the native hydrated state. (b) Surface cell-film method was used in c o m b i n a tion with positive or negative staining and related variants of our original technique (16) for spreading mitochondria (either as isolated or in situ) on a liquid surface to form an ultrathin layer. This can be picked up directly on a specimen grid and examined directly or after suitable staining and shadowcasting. The best results were obtained by using the valuable modification introduced by D. Parsons (55, 56) which embodies the advantages of negative staining-embedding. This simple method was carried out by inserting a clean glass needle or a freshly cleaved mica surface into the specimen of native tissue or the mitochondrial suspension without prior fixation. The needle or mica surface coated with the specimen was then dipped slowly into a 1 to 2 per cent sodium- or potassium phosphotnngstate solution at pH 7.2, preferably cooled to 0 to 4 , and ultrafiltered (using 10 m/~ Millipore filters). The mitochondrial membrane specimen spreads out into an exquisitely preserved ultrathin layer floating on the liquid surface. This coherent surface film (only a few hundred A thick and even less in certain areas) is picked up directly on a thin Formvar specimen support, carrying with it a tenuous phosphotungstate layer. U p o n drying (preferably in a cold chamber at 0 4 ) , this phosphotungstate film forms an amorphous embedding matrix of high electron opacity, and provides both protection against surface tension artifacts and excellent negative contrast. These spread-cell preparations are ideally suited for high resolution electron microscopy of mitochondrial membranes and related lamellar systems, provided the specimens are examined immediately under appropriate conditions to reduce irradiation damage and contamination. Thus, remarkable integrity of the fine structure of the membrane system of an entire mitochondrion is obtained which compares favorably with that observed in the best ultrathin
sections. This simple technique is susceptible of further refinement by spreading the specimens against constant pressure, spreading on undercooled liquid surfaces, or on clean mercury surfaces, etc. Also, by spreading the native mitochondria on solutions of buffered formalin, glutaraldehyde, osmium tetroxide, or potassium dichromate (0.1 to 2 per cent) the effects of prior fixation can be studied in detail. (c) Microdroplet cross-spraying techniques (23, 24) were used routinely in the study of mitochondrial membranes and isolated particles requiring minimal exposure to reagents, and a high degree of reproducibility. With the use of a special multiple-spraying device with suitably arranged separate capillaries for specimens and reagents, it is possible to obtain controlled, brief interaction of microdroplets of the specimen with microdroplets of 1 to 2 per cent potassium phosphotungstate at p H 7.2, uranyl acetate, or other heavy metal solutions. The cross-sprayed microdroplets collide and interact very rapidly shortly before impinging on the specimen grid. The twofold advantage is hereby gained of imposing predetermined spatial and temporal constraints on the specimen-reagent system during a highly reproducible preparative process. This method has to be contrasted with the standard negative staining procedure by which the PTA reagent is mixed with the specimen, and can produce variable degrees of modification before being sprayed and dried on the support. Moreover, as a result of the small size and relatively high speed of the cross-sprayed microdroplets, there is a favorable concurrence of rapid specimen cooling and limited drying to achieve a unique degree of preservation of the most labile structures (24). Finally, this technique is easily carried out and permits rapid processing of a large number of specimens. I n the present studies it was used for examination of all critical specimens, and to investigate the effects of fixation agents and heavy metal solutions on the fine structure of native mitochondrial components. Interesting results were obtained in the study of isolated EP (Fig. 20) by combining positive or negative staining of cross-sprayed microdroplets with subsequent shadow-casting at an extremely low shadowing angle (using platinum, or platinum-carbon
]hO~tE 4 Negatively stained beef heart mitochondrion; isolated in 0.5 M sucrose, and prepared by surface spreading on 1 per cent potassium phosphotungstate at pH 7.2 without prior fixation. Partly intact whole mount of flattened mitochondrial membranes showing regular arrangement of repeating particulate components. )< 6~,000. I~G~mE 5 Profile view of cristae with arrays of elementary particles (EP) in enlarged segment of Fig. 4. Polyhedral head pieces of the elementary particles are attached by stalks to the continuous dense outer layers of the cristae. X 4~0,000.
70
Tn~ JOURNAL OF CELL BIOLOGY VOLUME ~ , 1964
FERN.~.NDEz-MoR~.N ET
AL.
MacromolecularRepeating Unit of Mitochondrial Structure
71
evaporation in a vacuum system with liquid nitrogen cold-trap). The resulting reproducible "surface decoration" patterns may provide information on the underlying subunit structure of isolated EP. For attainment of highest resolution the carboncoated, fenestrated specimen films (29) were used without the normal ultrathin film supports (cast from 0.05 per cent Formvar solutions in ultrafiltered ethylene chloride). Instead, extremely thin embedding supports bridging the holes of the fenestrated fihn (ca. 100 to 1000 A in diameter) were obtained by the addition of very dilute (0.01 to 0.05 per cent) sodium silicate, gelatin, or gum arabic solutions to the specimen, preferably applied through microdroplet cross-spraying (24).
2. Thin Sectioning Techniques

In addition to the standard osmium fixation or glutaraldehyde fixation procedure followed by methacrylate or Araldite embedding and thin sectioning, the following two variants proved useful in the study of mitochondrial fine structure: a. Negative staining--embedding of ultrathin sections. Fresh tissues or mitochondrial suspensions (either unfixed and dried, or partially fixed by formalin vapor, or by ultraviolet irradiation) were sectioned without embedding by a diamond knife mounted on a Moran-Leitz ultramicrotome. When these coherent ultrathin sections were collected directly on a 1 to 2 per cent buffered phosphotungstate solution (pH 7) and picked up on filmed specimen grids, the fine structure of mitochondria and other lamellar systems showed up in striking negative contrast images. This favorable result is simply due to permeation of the "unsupported" ultrathin section by the "phosphotungstate glass" which acts as an "embedding" medium of high density and permits enhanced contrast and resolution. Despite its inherent limitations (mainly caused by drying artifacts), this interesting and simple approach has proved to be of value as a control procedure for the evaluation of negative staining in sectioned material. The results will be reported in a separate publication.
b. Ultrathin frozen sections of fresh tissues (24) prepared by cutting the frozen tissue with a diamond knife in a cryostat (at -- 30 to -- 180C) can also be examined directly without thawing by embedding the sections in vitrified heavy metal layers. A liquid nitrogen cooling device and appropriate low temperature methods of electron microscopy are required to implement this technique. However, numerous technical problems still remain to be solved before this method can be used routinely. For thin sectioning of suspensions and pellets of isolated mitochondria and submitochondrial fractions (ETP), the technique described earlier (29 a) for examination of oriented Tobacco Mosaic Virus gels proved very useful. The pellets or suspensions were sucked into thin plastic capillaries (0.1 to 0.2 m m internal diameter). The specimen capillaries were cut into short segments which could then be stained, dehydrated, and embedded. Extraction and dehydration or embedding artifacts were considerably reduced, particularly when using low temperature preparation techniques. c. Cryofixation and related low temperature techniques (19-23), including low-temperature dehydration and embedding in methacrylate after standard osmium fixation, were used primarily for control purposes in these studies. ELECTRON MICI~OSCOPY : A Siemens Ehniskop I was used, operating mainly at 80 kv (also at 40, 60, and 100 kv for selected specimens). Improved pointed filaments (20) of single-crystal tungsten with a tip radius of 1 to 10 # were used routinely with the double condenser system of the Elmiskop to provide intense microbeam illumination of high coherence and low angular divergence. With this arrangement, the image brightness (with a 2 to 10 /~ spot size, beam current 2 to 8 /za at 80 kv) was adequate for direct observation of image detail at the highest electron optical magnifications. Using clean multiple objective apertures of copper or platinum (50 #), the astigmatism of the objective lens could be readily corrected under direct observation and reduced to less than 0.2 micron; for high resolution studies, the measured
I~GURE 6 Isolated beef heart mitochondrion embedded in thin phosphotungstate (PTA) layer. Prepared by microdroplet cross-spraying procedure involving only brief interaction of mitochondrial suspension in 0.5 m sucrose with 1 per cent PTA. Notice characteristics paired arrays of elementary particles in profiles of fragmented cristae, which are readily distinguishable from the envelope of the mitochondrion. X 120,000. FtGURE 7 Enlarged segment of crista in a specimen similar to that shown in Fig. 6. Demonstration of three parts of the elementary particle: head piece, stalk, base piece. Invariant association of the arrays of head pieces EP with the underlying dense layer of crista. Segmentation and knob-like proturherance of dense layer at point of attachment of stalk. X 600,000.
72
THE JO~I~NAL Or" CELL BIOLOGY VOLU~E 2~, 1964
FERN.~,NDEz-MoR~.N ET AL.
73
astigmatism of the objective lens was 0.1 # or less. The higher efficiency of the pointed filament enabled us to use special Ilford High Resolution Plates as a routine (8 to 15 seconds exposure at 80,000 to 100,000 electron optical magnifications) for recording satisfactory electron micrographs which can be subsequently enlarged to yield higher useful magnifications than ordinary plates (22, 24). Conversely, it was also possible to obtain microbeams (0.5 to 2 /~ in diameter) of extremely low intensity by using the new pointed filaments under appropriate conditions, and by providing the double condenser system (CII) with apertures of 50 to 100 /~. This arrangement proved to be essential in examining certain highly labile membrane constituents. I n fact, useful micrographs from these specimens could be recorded only on high-speed emulsions (e.g. 35 m m Tri X Kodak, sensitized with gold thiocyanate solutions) by first focusing on adjacent specimen areas, and then rapidly shifting the low-intensity microbeam to the preseleered site. Irradiation damage and specimen contamination could be considerably reduced by using improved specimen cooling devices (Leisegang liquid nitrogen stage with special shielding apertures) in combination with low-intensity electron optics (19-24). For all critical measurements, calibration of the microscope was carried out at the time of recording the electron mlcrographs, using a diffraction grating replica carefully adjusted for the same specimen position. Ferritin molecules of uniform size (118 to 120 A diameter), prepared by density gradient ultracentrifugation and used b y J . W. Anderegg and F. Fischback for x-ray scattering studies, were frequently used as an internal calibration reference for certain specimens. However, in all of these experiments particular care was taken to guard against possible contamination with apoferritin or modified ferritin particles by running parallel controls without addition of ferritin or other calibration markers. The observations described here are based on the evaluation of more t h a n 3,000 plates and films, in which an average resolution of 10 to 20 A was consistently achieved.
BIOCHEMICAL
METHODS
The companion paper by Blair et al. (5) gives a full description of the methods of assay of the enzymic activity of mitochondria and submitoehondrial particles and of the analytical methods used in quantifying the components in these particles. RESULTS
Electron Microscopy of Isolated Mitochondria

Isolated beef h e a r t m i t o c h o n d r i a p r e p a r e d as described in the section o n M e t h o d s showed good preservation of form a n d fine structure. R o u t i n e checks were m a d e for c o n t a m i n a t i o n of mitochondrial preparations with myofilaments, ribosomes, a n d other extraneous particulate material by systematic e x a m i n a t i o n of pellets of isolated mitochondria. None of these c o n t a m i n a n t s were found in our s t a n d a r d preparations. Since m i t o c h o n d r i a are essentially fluid-filled vessels w i t h a n involuted internal m e m b r a n e syst e m of a few molecules in thickness, they are, in principle, well suited for direct e x a m i n a t i o n without sectioning. Once the internal fluid has been removed, the collapsed a n d distended mitochondrial m e m b r a n e s c a n be p e r m e a t e d b y buffered solutions of phosphotungstate a n d supported b y negative staining embedding. T h e e m b e d d e d layers are thin e n o u g h for high-resolution electron microscopy. Disruption, dissociation, a n d spurious rearr a n g e m e n t s of the delicate m e m b r a n e structures must be t a k e n into account as possible causes of artifacts. Various methods of negative staining were used to control the possibility of artifact formation. W e have h a d no indication t h a t the structures w h i c h are the principal subject of the present c o m m u n i c a t i o n are products of a particular type of preparative procedure. O n the con-
Fmv~v. 8 Enlarged portion of crista from specimen shown in Fig. 4. Notice characteristic knob-like structure of the base piece at the insertion point of each EP stalk. The base piece subunits and underlying regular segmented arrangement of dense layer discernible in both profile and surface view of crista. )< 600,000. FIGURE 9 Segment of negatively stained crlsta in a PTA spread-cell preparation of beef heart mitoehondrion. The regular know structures demarcating the base piece of each elementary particle form an integral part of the dense layer of the crista. Penetration of PTA between two apposed dense layers of the crista may account for the central dense band of the "mesozone" region. X 650,000.
74
THE JOURNAL OF CELL BIOLOGY VOLUME g~, 1964
FERN.~NDEZ-MORkN ET AL.
Macromolecular Repeating Unit of Mitoctwndrial Structure
75
trary, a wide range of preparative procedures has been used successfully for the visualization of these structures. In view of the large number of electron micrographs presented in this communication, we have prepared an Index summarizing all the essential structural items to be discussed in the section on Results and the corresponding figures to be examined in these contexts.
FINE STRUCTURE IN OF ISOLATED M1TO-
Preparations fixed with osmium tetroxide and dehydrated and embedded at low temperatures (Figs. 1 to 3) reveal a degree of preservation of fine structure of the membranes which compares favorably with that found in the best preparations of retinal mitochondria (22-25). In both types of preparations, the electron opaque layers of the cristae show characteristic granular or globular subunits about 70 to 80 A in diameter. This particulate subunit structure of the mitochondrial membranes is also regularly encountered in sectioned preparations fixed with glutaraldehyde or formalin and stained with uranyl or lead acetate. Apposing membranes of adjacent mitochondria (in sections of pellets) exhibited a regular pattern with a period of 100 to 160 A. Such patterns were first described by Pease (57). The typical angular configurations of the cristae, previously noted by Revel et al. (59), are particularly noticeable in preparations treated at low temperatures. There is a formal resemblance between the cristae in osmium-fixed and negatively stained preparations in respect to the dense layers. In both cases, the structured elements are localized in these two layers and it is only in the electron opaque layers that a regularity of structural pattern is recognizable. The possibility must be borne in mind that the material present in the particles associated with the cristae in the negatively stained preparations may become incorporated into the thicker electron opaque layers of the osmium-fixed preparations. THE ELEMENTARY PARTICLES: In whole membrane mounts of negatively stained mitoCHONDRIA SECTIONS:
chondria, all of the visible membrane surfaces appear to be covered with uniform particles having diameters the order of 80 to 100 A (cf. Figs. 4 to 7). The vast numbers of these particles--as well as their regularity and periodicity--are the most re markable features. W h e n first observed (23, 25), this repeating structural entity was designated the "elementary particle" (EP). In typical negatively stained preparations (Figs. 4 to 10), the elementary particles are arranged in recurring arrays associated with the cristae and with the inner membrane of the external mitochondrial envelope. Examination of whole-mount specimens reveals approximately 2,000 to 4,000 particles per square micron. If a fairly uniform distribution is assumed, the total number would be of the order of 10,000 to 100,000 elementary particles per mitochondrion depending on the size and type of mitochondrion. The average distance separating two adjacent particles was found to be about 110 to 115 A. If the particles are assumed to be spheres, 100 to 120 A in diameter, they would account for 10 to 15 per cent of the total mitochondrial volume. Similar estimates have been made by Smith (70, 71) for the volume occupied by these particles in insect sarcosomes. As seen best in paired arrays along a crista, each elementary particle consists of the following three recognizable components: a head piece, a stern or stalk, and a base piece which is the region of attachment of the stalk to the crista. The head piece is the most conspicuous portion, being polyhedral and asymmetric by virtue of the stem region (Figs. 5 to 12, 13 a to c). The measured diameter in micrographs of freshly prepared specimens is about 80 to 100 A (cf. Figs. 4 to 12). The ratio between the long and short axes of the asymmetric head piece comes to values of 1.2 to 1.4. W h e n mitochondria are fixed with aldehydes (formaldehyde, glutaraldehyde), and the resulting preparation treated with iodine or dichromate, the head often appears surrounded by a " h a l o " of granular components (20 to 30 A diameter) delimiting the periphery.
FIGVRE 10 Electron micrograph of negatively stained beef heart mitochondrial membranes recorded with low-intensity microbeam illumination, and specimen cooling ( - 8 0 ) to reduce irradiation damage. Under these favorable conditions, the dense substructure of the EP head pieces can be more frequently observed in suitably oriented specimens. Notice larger size of detached EP. Specimen cooled and examined immediately after preparation by cross-spraying technique. M 860,000.
76
THE JOm~AL OF CELL BIOLOGY VOLUM~ ~ , 1964
FERN~NDEZ-MOR,~N ET AL.
77
The head piece exhibits a characteristic, electron opaque core which can be demonstrated consistently, providing special precautions are taken (Figs. 10 to 12,. 13 a to d). The visualization of this electron-opaque region of the head is influenced by exposure of the specimen to the intense electron beam. The apparent fading away of the dense central region while the head piece is being observed with electron optics at high magnification has been repeatedly noted. The head core is therefore best recorded when low-intensity, microbeam illumination is used at relatively low electron optical magnifications ( 10,000 to 20,000). For observations at higher electron optical magnifications ( X 40,000 to 80,000), the stage containing the specimen is cooled with liquid nitrogen (19, 23-26) to avoid deterioration and fading of the image. As demonstrated in through-focus series (Figs. 11 a, b and 12 a, b), the electron opaque core of the head piece is of variable width (ca. 20 to 60 A). The core shows up best against a low contrast background such as is achieved by embedding the specimen in dilute (0.1 to 0.3 per cent) phosphotungstate solutions. In many ways the electron opaque core of the head piece resembles the corresponding electron opaque core seen in electron micrographs of cytochrome oxidase (annular forms of the complex). The electron opaque core has been consistently observed also in preparations treated with uranyl acetate or dichromate and in specimens fixed with formalin or glutaraldehyde (Fig. 13 d). The stem or stalk region appears to have variable size and configuration depending on the type of negative staining. W h e n the specimen is prepared by microdroplet cross-spraying with a very brief period of interaction with phosphotungstate, the stalk seems to be very short (Fig. 10) and the head piece seems to be wedged in di-
rectly within the mesoregion of the crista. In these cases, the elementary particles assume a teardrop formation (cf. Fig. 7). In spread cell preparations, the stalk is seen to be well developed (Figs. 5, 8) and is approximately 20 to 40 A wide and 40 to 50 A long. At the point of attachment of the stalk to the crista a characteristic knob-like formation, approximately 40 to 50 A in diameter, is regularly seen. An invariant feature of all the electron micrographs examined is the correlation between the arrays of head pieces of the elementary particle and the electron opaque layer of the crista to which the head pieces are attached by stalks (cf. Figs. 7 to 9, and 15). We consider the dense layer to be an expression of the end-to-end alignment of the base pieces of the elementary particles. The segmentation of this dense layer in the form of recurring knobs parallels the periodicity of the head pieces and the stalks. Since the average distance between the head pieces is about 115 A, the same distance must be assigned to the width of a base piece if the dense layer is assumed to be made up of fused base pieces. T h e thickness of the dense layer is 40 to 50 A. The third dimension of the base piece would have to be 115 A to be consistent with the assumption of a continuous layer. ELEMENTARY PARTICLES IN SUBMITOCHONDRIAL FRAGMENTS: The same type of repeating particulate structure is regularly observed in submitochondrial fragments (ETP) obtained by sonication (@ Figs. 14 and 15). The particles are an intrinsic feature of the mitochondrial membranes even when the mitochondrion has been disrupted. In fact, the presence of elementary particles both in intact mitochondrial membranes and in mitochondrial fractions was the starting point for the biochemical isolation. The three parts of the elementary particle (head
FramES l l a and b, and 12 a and b Electron micrographs selected from through-focus series of negatively stained cristae prepared and examined under same experimental conditions as described in legend for Fig. 11. Even under such favorable conditions, the electron opaque core with indications of substructure can be reproducibly demonstrated only in certain EP head pieces (arrows). Figs. 11 a, b, X 750,000; Figs. 12 a, b, X 450,000. FIGURES 13 a to d Inset enlargements showing electron opaque core in head pieces of selected elementary particles. Experimental conditions as described in legends for Figs. 10 and 11. Figs. 13 a to c, specimens stained with 1 per cent potassium phosphotungstate without prior fixation. Fig. 18 d, specimen fixed with 2 per cent glutaraldehyde at pH 7.2 prior to staining with 1 per c~nt uranylacetate. Magnifications: Fig. 18 a, X 1,250,000 Figs. 13 b to d, )< 950,000.
78
THE JOURNAL OF CELL BIOLOGY VOLUME 22, ]964
FERNANDEz-MoR~N ET AL.
79
piece, stalk, and base piece) have the same relative dimensions in the submitochondrial fragments as they do in the intact mitochondrion. We have frequently used ferritin as a reference particle for estimation of size. O n the basis of careful measurement of the diameters of specially graded ferritin particles which show up as spherical units with an electron-opaque core, we have assigned a value of 118 A which is in good agreement with the unpublished x-ray data of J. W. Anderegg and F. Fischback. The head piece of the elementary particle is remarkably similar to ferritin, although the average diameter is slightly less than that of ferritin and the electron opaque core is less pronounced. In m a n y preparations of mitochondria and submitochondrial fractions, completely detached particles are observed near the membranes. If these particles are, in fact, detached elementary particles, their size would be significantly larger than that of the head piece (approximately 120 to 140 A in diameter compared to 80 to 100 A). These detached particles are larger than those of ferritin.
VISUALIZATION PARTICLE BY OF THE ELEMENTARY TECHNIQUES: ALTERNATIVE
mitochondrial specimens in vacuum-tight microchambers (20, 22-24) and related techniques. Visualization of the subunit structure was not significantly affected by such variations. Positive results were also obtained in the examination of specimens prepared by ultrathin sectioning after cryofixation, by negative staining combined with ultrathin sectioning, and by embedding in gelatin and other water-soluble media. These results will be described in detail in a separate communication. BIOCHEMICAL
OF
RESULTS
Mitochondria (in situ and after isolation) have been fixed and stained with a variety of reagents (including buffered solutions of osmium tetroxide, formaldehyde, glutaraldehyde, potassium permanganate, uranyl acetate, potassium dichromate, silicotungstate, potassium iodide, sodium bromide, and cadmium iodide). Examination of such treated preparations has consistently revealed essentially the same type of particulate structure. The following types of variation in the experimental conditions were also explored: interaction of the mitochondrion with the fixing reagent prior to staining for different times and at various temperatures; cooling of the specimen during observation; electron microscopy of partially hydrated
ELEMENTARY PARTICLE AND THE UNIT ELECTRON TRANSFER: Full evidence has already been presented (5) that the components of the electron transfer chain as well as electron transfer activity (as measured by succinoxidase and D N P H oxidase activities) are concentrated some 2.6 times during isolation of the elementary particle from mitochondria. I n presenting the results that follow, we shall adopt the conclusion of Blair et al. (5) that the elementary particle is, in fact, the intact electron transfer chain stripped completely from the primary dehydrogenase complexes and largely, if not completely, from the structural protein-lipid network. We shall now consider the experimental evidence that bears on the molecular weight of the isolated elementary particle and on its probable molecular dimensions. This information is essential to answer the question of the identity of the isolated elementary particle with the counterpart particle seen in the cristae of intact mitochondria.
MOLECULAR TARY PARTICLE WEIGHT FROM OF THE ELEMENDATA: ANALYTICAL
The cytochrome a content of the elementary particle isolated by processing through two cycles of exposure to the cholate-deoxycholate mixture, followed by fractionation with a m m o n i u m sulfate,
I~GVRE 14 Electron micrograph of negatively stained preparation of the electron transfer particle ( E T P ) ~ a submitochondrial fraction. Elementary particle arrays closely resemble those in intact cristae; average dimensions of head piece and other components of EP same as those for corresponding structures in EP of intact mitochondria; apposition of two dense layers to form a single fused central layer. Specimen preparation as described in legend for Fig. 6. X 600,000. l~eulm 15 Negatively stained segment of erista from isolated whole Initochondrion (enlargement of Fig. 6). Specimen preparation conditions are essentially the same as for specimen shown in Fig. 14. Note identity of form and arrangement of elementary particles in erista of intact mitoehondrion and in the electron transfer particle (ETP). X 660,000.
80
TIlE JOUENAL OF CELL BIOLOGY VOLUME ~ ,
1964
FEaN,(NDEZ-MoR~.N ET AL.
81
TABLE III Minimal Molecular Weights of the Four Complexes of the Electron Transfer Chain at the Highest Purity Levels Achieved
Molcular weight Complex Designation Protein basis Lipoprotein (3 ~0 lipid)
II III IV Total
DPNH-coenzyme Q reductase* Succinic-coenzyme Q reduet ase ~ QH~-cytoehrome c reduetase Cytoehrome oxidase][
550,000 208,000 200,000 426,000 1,384,000
786,000 297,000 285,000 609,000 1,976,00()
* Unpublished data of J. Merola. Data of Ziegler and Doeg (81). Data of Rieske, Zaugg, and Wharton (61). ][ Data of Criddle, Bock, Green, and Tisdale (13); Ambe and Venketaraman (l).
FIGURE 16 Sedimentation diagram of an elementary particle preparation. The particles were suspended in 90 per cent D20 which was 0.0~ M in Tris chloride and 0.001 I~ in EDTA. Protein concentration, 11 mg per ml. Rotor speed, 50,700 arM; rotor temperature, 6. Time interval between photographs, 4 minutes. The apparent sedimentation constant at 6 is 13.6 S~0.,~ = 5~. The particles were solubilized with phosphotungstate as described in the text. The slow minor peak in the diagram is that of phosphotungstate. The sedimentation diagrams read from right to left. is 4.2 m g m o l e s / m g of protein. O n the basis of 6 molecules of cytochrome a per molecule of elementary particle (41), the minimal molecular weight of the particle would be 6 4.2106 1.4 X tion of the four component complexes under specified conditions (30, 44, 45). In the companion paper by Blair et al. (5), we have discussed the evidence that the four complexes combine in 1 : 1 : 1 : 1 molecular stoichiometry. Table I I I contains the minimal molecular weights of each of the four complexes as determined by chemical analysis. The four complexes used in the reconstitution studies were at the stage of purity indicated in Table III. As we shall discuss later, complex I at the purity level used in these experiments is still highly impure, but the other three complexes are probably close to the limit of purification. The sum of the minimal molecular weights of the four complexes comes to 1.38 X 106 on a protein basis--a value which is in good agreement with the minimal molecular weight calculated on the basis of the composition of the isolated elementary particle. Since the four complexes may combine in 1 : 1 : 1 : 1 molecular stoichometry, it is necessary to
106 on a protein basis. Cytochrome a is selected for this computation since its concentration is more accurately determined than any other component of the chain. The minimal molecular weight is not necessarily the actual molecular weight. In a unit of minimal molecular weight, 1.4 X 106, it is assumed that there would be one molecule each of cytochrome q, succinic dehydrogenase (fs), and D P N H dehydrogenase (fD). The actual unit of electron transfer could contain more than one molecule of these components. MOLECULAR WEIGHT OF THE ELEMENTARY P A R T I C L E F R O M T H E D A T A OF R E CONSTITUTION tary EXPERIMENTS: The elemen-
particle has been reconstituted by recombina-
82
THE JOURNAL OF CELL BIOLOGY VOLUME ~ ,
1964
establish for only one of them that the minimal and actual molecular weights are the same in order to extrapolate to the actual molecular weight of the reconstituted unit. The molecular weight of the QH2-cytochrome c reductase (complex III) was determined by sedimentation analysis (J. Rieske and P. Yang, unpublished data) and found to conform to the minimal molecular weight listed in Table III. Takemori et al. (73) estimated the molecular weight of cytochrome oxidase (corn-
plex IV) to be about 500,000--a value in good agreement with that for the minimal molecular weight of the complex (615,000 for the lipoprotein complex). Singer, Kearney, and Massey (66) have isolated a form of succinic dehydrogenase almost identical in composition with that of complex II. T h e molecular weight which they determined by sedimentation analysis of their preparation came to a b o u t 250,000--in good agreement with the value for the minimal molecular weight of comIV
TABLE
Inhibition of Electron Transfer Activities by Sodium Phosphotungstate

Concentration of sodium phospho* tungstate
Activity measured
Particle type
Inhibition
per cent
1. Oxidation of D P N H by 0 2
Mitoehondria (heavy)
Mitochondria (washed with 0.9% KC1) ETP~
5 4 8 4 5.5 6.5 2.0 6.0
X X X X X X X X
10-7 10-.-6 10_6 10_5 10-5 10-4 10-5 10- 5
1.2 X 10- 4
EP
5.0 X 10-6 3.0 X 10- 5 5.5 X 10-5 5.0 4.0 7.0 2.0 5.0 8.5 5.0 2.0 4.0 1.0 3.5 5.5 X X X X X X >( X X X X X 10- 7 10-6 10--6 10- 6 10- 6 10- 6 10-7 10- 6 10-6 10- 5 10- 5 10-5
I0 50 75 10 50 75 10 50 75 10 50 75 I0 50 75 10 50 75 10 50 75 10 50 75
2. Oxidation of succinate by 0 2
Mitochondria (heavy)
Mitochondria (washed with 0.9% KCI) ETPH
EP
The above experiments were all carried out at a particle protein concentration ofS0 #g per assay. When the protein concentration is increased to 5000 #g per assay, the concentration of sodium phosphotungstate required to achieve the same degree of inhibition as at the 50 #g protein level is 3 to 5 times higher. When specimens are prepared for negative staining, the usual concentration range of sodium phosphotungstate concentrations is 0.1 to 1%, which corresponds in molar terms to 3.3 )< 10- 5 ~ -- 3.3 )< 10-4 M. Mitochondria or submitochondrial particles which have been exposed to phosphotungstate at inhibitory concentrations can be washed free of the reagent; the electron transfer activity after washing is the same as the original, uninhibited activity.
FERN.i..NDEZ-MOR.~I ET At,. MacromolecularRepeating Unlt of Mitochondrial Structure
83
plex II. The available evidence supports the view that the minimal molecular weight calculated from analytical data is, in fact, the actual molecular weight of the elementary particle.
SEDIMENTATION MENTARY ANALYSIS OF THE ELEPARTICLE : The isolated elementary particle shows a strong tendency to aggregate despite the presence of residual deoxycholate which is not removed by exhaustive washing in 0.25 sucrose. Prolonged sonication partially disperses the suspension, but neither permanently nor completely, to the stage of the monomeric species. This tendency of the particles to aggregate has effectively blocked our attempts to establish the homogeneity of the preparation and to determine its molecular weight by sedimentation analysis. The ultracentrifugal analysis of suspensions of the elementary particle following dispersion by sonic irradiation shows a broad distribution of particle sizes. We have sorted out the particles from the most aggregated to the least aggregated in successive fractions collected in the trunnion head of the Spinco ultracentrifuge and have determined the composition of each of these fractions No measurable differences were found in the chemical composition of the different fractions. Thus, it is clear that the heterogeneity of particles is an expression of differences in degree of physical aggregation, but not in chemical composition, The elementary particles can be brought into true solution by interaction with phosphotungstate. The optimal level for "solubilization" is a l : l ratio by weight of elementary particle (as protein) and phosphotungstate, The solution of the elementary particle-phosphotungstate complex is centrifuged at 50,000 RPM for 15 minutes. The precipitate consists of two layers--a hard packed sediment which is discarded and a fluffy layer which is mixed with 0.01 ~a Tris buffer of pH 7.5. Sedimentation analysis, by Dr. Pauline Yang, of
this phosphotungstate complex of the elementary particle showed a major component with a weight average sedimentation coefficient of 52 X 10- ~ see. after extrapolation to zero concentration Fig. 16). Chemical analysis of the complex showed that 1 gm of elementary particle protein was combined with 0,9 gm of lipid and 1 gm of phosphotungstate. The minimal molecular weight of this complex should, therefore, be 2.9 X 1.6 X 106 = 4.6 X 106 (the minimal molecular weight of this particular preparation of EP was 1.6 X l0 p on a protein basis). The observed molecular weight determined by the modified Archibald approach to sedimentation equilibrium was 5.8 X 106, in fair agreement with the minimal molecular weight calculated from analytical data. The difference could be accounted for by water of hydration. From the experimentally determined values for the partial specific volume (0.727 cm~/gm) and the sedimentation coefficient (52 X 10-13 sec.), the frictional coefficient of the elementary particlephosphotungstate complex can be calculated by the Svedberg equation:
(cf.
M(l
- - ~0)
/-
NS
where f is the frictional coefficient of the elementary particle complex; the molecular weight of the complex (5.8 X 106 gin/mole); 5, the partial specific volume; p, the density of the solvent (1:00 gm/cm3); S, the sedimentation coefficient; and N, Avogadro's number. W h e n the values indicated are substituted in the equation, the value for f comes to 5.06 X 10-7 (C.G.S. units). The frictional coefficient of a spherical molecule of the same molecular weight was calculated by the following equation:
M,
:o = 6~, @ I ~ : /
(3Mg, ~l/a
FIGURE 17 Electron micrograph of isolated elementary particles negatively stained with 1 per cent potassium phosphotungstate by microdroplet cross-spray technique. Isolation of elementary particles as described in text; electron transfer activity of particles fully preserved. In this typical field, discrete and conglomerated particles are seen in different orientations; ferritin molecules (dense core, diameter 1~0 A) added as a marker. X 4~0,000. FIGURE 18 Preparation of isolated elementary particles embedded in thin film of phosphotungstate, selected from another specimen series. Note: asymmetry of particles; relative uniformity of particle size (1~0)< 160 A); minimal aggregation in this particular specimen. )< 330,000.
84
THE JOURNALOF CBLI~BIOLOGY VOLUME~ , 1964
FEnN~,NDEz-MoR.~N E'r .eL.
85
Fmua~ 19 Isolated elementary particles examined by the shadow-casting technique. Details of preparation as in legend for Fig. 17. Individual particles of variable size and polyhedral shape clearly visible. X 300,000. FtouaE ~0 Microdroplet of isolated elementary particles, preparation positively stained with uranyl acetare and shadowed with platinum-carbon at very low grazing angle; the resulting "surface decoration" patterns indicate possible underlying subunit structure of individual elementary particles. X 600,000.
86
T~
JOURNAL OF CELL BIOLOGY ' VOLUMEg~, 1964
where y is the viscosity of the solvent (1.47 X 10-~ poise) and the other symbols have been defined above. The value for fo comes to 3.28 X 10-7 when the indicated substitutions are made. Therefore,
f fo 5.06 X 10 -7 3.28 X 10 -7
made for lipid and phosphotungstate). The discrepancy between the observed and calculated molecular weights provided independent confirmation of the asymmetric character of the iso~ lated elementary particle and of its soluble phosphotungstate complex.
ELECTRON MICROSCOPY PARTICLES : OF The ISOLATED
1.54
The value for the ratio of f to fo clearly indicates that the elementary particle-phosphotungstate complex is asymmetric, but the degree of asymmetry cannot be evaluated in terms of axial ratio until the hydration of the complex can be specified with precision.
EFFECT ENZYMIC OF ACTIVITY PHOSPHOTUNGSTATE : ON
Phosphotungstate, at the levels used for the preparation of samples to be examined in the electron microscope, is a relatively mild reagent in its effects on enzymatic activity of mitochondria and submitochondrial particles (cf. Table IV). Even when enzymatic activity was almost completely blocked, the effects were found to be reversible. Thus, after exposure to inhibitory levels of phosphotungstate, the particles regained full activity when washed free of the reagent. These experiments clearly indicate that treatment with phosphotungstate under the described conditions does not lead to irreversible structural and functional modifications. Any such effects would be inconsistent with the partial inhibitory action of phosphotungstate at concentration levels that are used in electron microscopy and with the reversibility of whatever inhibition is observed at these levels. X-RAY SCATTERING DATA: Drs. J. A. Anderegg and N. Chonacky have examined preparations of the elementary particle and of the phosphotungstate complex for low angle x-ray scattcring. The Guinier plot (the log of the intensity of scattering versus the square of the scattering angle) was not linear--an indication that aggregation of particles took place during the period of the measurements. The available data clearly indicate that the smallest particles in the preparation have a radius of gyration of 60 A which corresponds to a diameter of 160 A. A spherical molecule, 160 A in diameter (density 1.2), corresponds to a molecular weight of 1.55 X l0 s. The preparations of the elementary particles submitted for examination would have a molecular weight in the range of 3 to 4 X 106 (after allowance was
preparations of isolated EP present inherent difficulties as to an examination by electron microscopy. Although special precautions were taken to examine the specimens as soon as possible (within a few hours of isolation), marked color changes, increasing opalescence, and other visible changes were noted as the preparation stood. The phase contrast microscope showed the appearance of minute particles (0.1 to 0.2/z) a few hours after preparation, even when the specimen was kept at 0C. The aggregation was markedly enhanced by agitation. In view of this remarkable lability, and of the unavoidable contamination with bile salts, it has been difficult to obtain unequivocal electron microscopic evidence of a homogeneous, uniform size of the isolated elementary particles. As seen in the best preparations (negative staining), the isolated EP's are polyhedral; the shape resembles that of a prolate ellipsoid. Depending on their orientation within the phosphotungstate film, the dimensions of the smaller diameter of the particles is 100 to 120 A whereas the dimension of the long axis is 140 to 180 A (Figs. 17, 18). All of these preparations, however, show characteristic electron opaque components in the core such as can be seen in the in situ particle. A number of preparations were examined by shadow casting in combination with positive staining (Figs. 19, 20). In these preaparations the individual particles stand out more clearly, although there is still a variation in size from about 120 to 180 A. Indications of subunit structure of approximately 20 to 30 A can be detected in shadowed specimens (Fig. 19) and in positively stained particles which are subsequently "decorated" by shadowing with platinum-carbon at an extremely low angle (Fig. 20). On the basis of electron lnicroscopy, it is, therefore, possible to establish the fact that the unit of electron transfer has a particulate character with a marked asymmetric configuration, the short and long dimensions being about 120 and 180 A, respectively. However, further improvements in preparative techniques will be needed to obtain
ELEMENTARY
FERN~NDEZ-)/[OR,~NET AL. Macromolecular Repeating Unit of Mitochondrial Structure
87
elementary particles of uniform size and shape comparable to that of other isolated multienzyme complexes (for example, the pyruvate dehydrogenase complex examined by H. Fernfindez-Morfin in collaboration with Lester Reed, 28, 58). ELECTRON MICROSCOPY OF THE REC~ONSTITUTED ELEMENTARY PARTICLE: In principle, the specimens of reconstituted EP are subject to the same inherent limitations as are the isolated EP's. However, the variability in size was less (predominance of 120 to 140 A units). Moreover, in preparations of reconstituted EP there was a marked tendency for aggregation of the particles into a regular arrangement, resembling paracrystalline arrays (Figs. 21, 22). I n some negatively stained preparations the phosphotungstate seems to exert a marked dissociating effect so that the reconstituted EP breaks down into smaller units. Nevertheless, these preparations afforded the opportunity of comparing by electron microscopy the dimensions and configurations of the reconstituted EP with those of the purified individual complexes. ELECTRON MICROSCOPY OF INDIVIDUAL COMPLEXES: Although most of the complexes were available in purified form, the presence of bile salts in the respective preparations of these complexes interfered with the interaction with phosphotungstate. ~l-he best preparations were attained by microdroplet cross-spraying involving very small volumes (microdroplets of the order of I /z or less) and very short interaction times with phosphotungstate or with uranyl acetate. An additional difficulty was encountered in trying to establish the precise delimitation of the zone of each individual complex. The most reproducible and clear-cut preparations were obtained with the largest of the four complexes, namely, cytochrome oxidase. As shown in Figs. 23 a, b, the cytochrome oxidase preparations appear as a mosaic of well defined units, approxi-
mately 60 to 80 A in diameter with an electronopaque core of approximately 20 to 30 A. As described earlier (23, 25), the electron-opaque components (heine, Cu) often feature a characteristic tetrad arrangement with a diameter of the order of 15 to 20 A units. This is surrounded by a " h a l o " to give a total diameter of about 60 to 80 A. Preparations of the other complexes (I, II, and I I I ) were not so well defined as preparations of cytochrome oxidase. However, complex I ( D P N H coenzyme Q reductase) was characterized by the presence of discrete, particulate units approximately 40 to 50 A in diameter (Fig. 26). Complex I I (succinic coenzyme Q-reductase) showed a marked tendency to aggregate. The average uniform particles were approximately 50 to 60 A in diameter (Fig. 25) and resembled those of complex I in appearance. Complex I I I (QH2-cytochrome c reductase) appeared as an electron opaque series of particles approximating 40 to 50 A in diameter (Fig. 24). Based on this preliminary data, it is possible to assume that the four individual complexes, as seen by electron microscopy, could be fitted into the isolated elementary particle. This is an important point which, however, requires further investigation. The availability of more highly purified complexes, and further improvements in high resolution electron microscopy of the individual complexes, and of the elementary particles, should eventually permit unequivocal determination of the actual dimensions. DISCUSSION
Validity of Structures Revealed by Negative Staining

Although the elementary particles can best be seen in negatively stained preparations of mitochondria without prior fixation (either in situ or after isolation), particles of essentially the same characteristics can be observed in electron micro-
FIGURE 21 Preparation of reconstituted elementary particles negatively stained with phosphotungstate; elementary particles reconstituted as described in the text. Specimen prepared as in legend for Fig. 17. Note: discrete asymmetric particles and aggregates of individual particles. Ferritin molecules (dense core, diameter 118 to 1~0 A) added as calibration markers. X 420,000. FIGURE 22 Reconstituted elementary particles negatively stained and prepared as in legend for Fig. 21. Note: discrete particles (120 to 140 A) and tendency to form extended linear aggregates resembling arrays of elementary particles as seen in ETP. X $80,000.
88
THE JOURNAL OF CELL BIOLOGY VOLUME 2~, 1964
:FERI~'DEz-MoR~N
ET AL.
89
graphs of mitochondria treated with fixatives and reagents other than phosphotungstate (23, 27). In the section on Results, mention is made of the wide variety of supplementary techniques that have been used for visualization of the elementary particle and the wide range of experimental conditions under which visualization can be achieved. The same type of structures have been reported in negatively stained preparations of mitochondria from several sources by Parsons (55, 56), Smith (70, 71), and Stoeckenius (72). The presence of a well defined particulate component directly associated with mitochondrial membranes in negatively stained preparations now seems to be established beyond dispute. The negative staining technique (7) has already disclosed the substructure of viruses, and proved to be extremely useful by yielding results that can profitably be compared with x-ray data and biochemical information to deduce the actual arrangement of the protein subunits in certain spherical virus particles (8). Although specimens embedded in a thin film of phosphotungstate are protected from artifacts due to surface tension forces, other artifact possibilities must also be considered. In particular, exposure of the unfixed labile lipoprotein systems of mitochondrial membrane to the phosphotungstate solution could conceivably lead to disruption, rearrangements, and other modifications of the original structure. It is, therefore, important that essentially the same type of structures can be observed in mitochondrial membranes which have been previously fixed with formalin, glutaraldehyde, osmium, and other reagents (27, 70, 71), provided that the necessary precautions are taken.
The Biochemical Inertness of Phosphotung-state

From the data presented in the section on Results (cf. Table IV), phosphotungstate at pH 7.0
does not markedly inhibit electron transfer activity (DPNH oxidase and succinic oxidase activities) at concentrations that are sufficiently high to prepare specimens for examination in the electron microscope. In fact, it has proved to be practically impossible to abolish completely enzymic activity even at the highest concentration of phosphotungstate normally used in the preparation of negatively stained specimens. Moreover, the inhibitory effect of phosphotungstate on isolated mitochondria is reversible since activity can be restored by washing the treated mitochondria by centrifugation and resuspension in a medium free of the reagent. The inference that phosphotungstate at concentration levels relevant to negative staining for electron microscopy has only a small, reversible effect on enzymic activity is highly significant; it would be difficult to invoke any major structural modification in the electron transfer chain at relevant concentration levels. The criterion of enzymic activity is even more stringent than is any physical measurement. Thus, the biochemical evidence is decisive in ruling out phosphotungstate as a reagent likely to induce chemical modifications in the structure of the mitochondrion. Phosphotungstate stands in contrast to osmium and other heavy metal reagents which at the concentration levels usually used in electron microscopy completely abolish enzymic activity and undoubtedly induce profound chemical and physical changes in mitochondria and other membranous systems. It should also be pointed out in this connection that negative staining with phosphotungstate and drying does not impair appreciably the viability of infectious polio virus and of certain other virus particles. Phosphotungstate is, in fact, being used as a reagent for depolymerizing aggregates of the elementary particle into the soluble monomeric species. For this purpose, it appears to be superior to any of the bile salts.
FtomtE 23 Electron micrographs of isolated cytochrome oxidase (complex IV) negatively stained with phosphotungstate; details as in legend for Fig. 21. Note: mosaic of repeating substructures of annular shape (diameter of 60 to 80 A) with electron-opaque core of 20 to 30 A. l~OtrRE 23 a, X 500,000; l~koURE28 b, X 700,000. EIaURES 24 to !~6 Electron micrographs of microdroplet preparations of individual complexes III, II, and I, respectively; stained with 1 per cent uranyl acetate by crossspray technique. Note: repeating substructures, about 30 to 60 A in diameter, with tendency to assume spherical or annular appearance. Figs. 24 and 26, X 500,000; Fig. 25, X 450,000.
90
THE JOURNALOF CELL BIOLOGY V o L ~ 2~, 1964
FERNkNDEz-MoR.~.N ET ~L.
91
FIouuEs 27 to 29 Comparison of the elementary particle as seen in electron mierographs of: native EP (Fig. 27); isolated EP (Fig. 28); reconstituted EP (Fig. 29). Details as in legend for Fig. 17. N 900,000.
Dimensions of the Elementary Particle in Situ

We are defining the elementary particle as a composite of three parts: (1) the roughly spherical head piece (80 to 100 A in diameter); (2) the cylindrical stalk (30 X 50 A); and (3) the base piece (40 112 A). There appears to be a precise
relation between the segmentation of the limiting membrane layer of the crista and the periodicity of the head piece. T h a t is to say, each head piece is linked by the stalk to a corresponding base piece. The base piece accounts for the segmentation of the outer layer of the membrane of the crista. It is difficult to specify where one base piece
92
THE JOURNAL OF CELL BIOLOOY VOLUME ~2, 1964
ends and the next begins. We shall make the simple assumption that the base piece can be no wider than the distance between two particles (i.e., ca. l l0 A) and that the dense layer of the membrane of the crista (diameter 40 A) is made up entirely of nesting base pieces. According to this interpretation, the crista is bounded by a structured, electron-opaque layer, 40 A in diameter, made up of fused base pieces to which the head pieces are joined by stalks. Within
mentary particle (head piece plus stalk plus base piece) should be sufficient to accommodate one complete electron transfer chain. This volume may be computed by assuming the head piece to be a sphere of diameter 90 A; the stalk a rectangular cylinder (50 X 32 114 A); and the base piece a cylinder (114 114 45 A). If we assume that there is an underestimation factor of about 12 per cent (observed size versus theoretical size) and correct each of the measured dimensions by a 12
TABLE V
Theoretical Molecular Weight of the Electron Transfer Chain Based on the Stoichiometry and Minimal Molecular Weights of the Protein Components
Contribution to molecular weight of the chain
Protein component
Molecular weight Reference
Molecules/ Chain
Succinie dehydrogenase (fs) D P N H dehydrogenase (fD) Cytoehrome a (a) Cytoehrome b (b) Cytochrome cl (cl) Cytochrome c (c) Non-berne iron (FeN~) Total
80,000* 37,5005 70,000 28,000 37,000 13,000 12,500/Fe
(81) (52) (1) (31) (6) (74) (60)
1 1 6 3 1 2 18
(5) (5) (5) (5) (5) (5) (5)
80,000 37,500 420,000 84,000 37,000 26,000 225,000 909,500
* Molecular weight of succinic-coenzyme Q reductase (81) corrected for 8 atoms of non-heine iron and one molecule of cytochrome b; 12,500 gm of protein per gram atom of non-heine iron and 28,000 gm of protein per mole of cytochrome b are assumed. ~:Molecular weight of D P N H flavoprotein (52) corrected for 7 atoms of non-heine iron, if 12,500 gm of protein per gram atom of iron are assumed. The assumption is being made that all the non-heine iron proteins have on the average the same molecular weight as the one isolated from complex I I I by Rieske et al. (60). this boundary layer (i.e. within the interior of the crista) are more irregular structures that are less electron opaque. The structural protein-lipid network is probably localized in these interior structures. The width of each crista is highly v a r i a b l e - so much so that it appears that the material in the interior may be capable of extrusion. In some segments of a crista the two outer layers are in apposition, whereas in other segments the two outer layers are separated by distances of 200 to 1,000 A. In such distended tubules, the interior is filled with a matrix-like material that is readily distinguishable from the electron-opaque outer membrane layer. The total volume occupied by a single eleper cent increment, the molecular weight contributions of head piece, stalk, and base piece (computed from the volumes with the assumption of 1.25 as the value for the density) are respectively 497,000, 193,000, and 607,000. The total computed molecular weight then comes to a value of 1.3 X l0 s. A sphere of density 1.25 with a molecular weight of 1.3 X l0 s would have a diameter of 148 A.
Dimensions
of
the
Isolated
Elementary
Particle In the companion paper, Blair et al. (5) described the preparation of the unit of electron transfer having a minimal molecular weight of
FERN,~NDEz-MoR~,N ET M~. Macromolecular Repeating Unit of Mitochondrial Structure
93
1.4 X 106 on a protein basis. We shall, for the moment, make no correction for the lipid content of the isolated particle. The minimal molecular weight is computed on the basis of the composition of the particle; the assumption is made that one molecule of succinic flavoprotein or of D P N H flavoprotein is present per molecular weight unit of the electron transfer chain. The ultracentrifugal data given in the section on Results indicate that the minimal calculated molecular weight and the actual determined molecular weight are sufficiently close to permit the assumption of identity. Following the initial studies of Blair et al., further progress has been made in the isolation procedure,
particle as isolated by the improved procedure of Tisdale, Tzagoloff, and Green (75). The isolated unit no longer shows the three distinct parts characteristic of the in situ elementary particle. W e presume that, when the elementary particle is detached from the mitochondrial membranes, the three component parts round up to form a compact unit in which the individual parts are no longer recognizable. The exact shape of the isolated particle is, therefore, not necessarily relevant to the question of the shape of the particle in situ. I n any event, the isolated particle must be "solubilized" by conversion into its phosphotungstate complex before
T A B L E VI
Theoretical Molecular Weights of the Complexes of the Electron Transfer Chain Based on the Composition of the Complexes and the Stoichiometry of the Protein Components
Computed molecular weight Complex I II III IV Total Composition* fD; (FeNn)s fs; (FeNn)8; b (b)2; (FeNn)2; ct (a)6 Reference (46, 52) (81, 60) (46, 60) (1, 13) (Protein basis) 137,500 208,000 118,000 420,000 883,500 (Normalized for 3 ~ lipid) 196,500 297,000 168,000 600,000 1,262,000
* See Table V for abbreviations of the components. and a unit of molecular weight 1 X 106 on a protein basis has been approached (75). If the molecular weight is normalized for the presence of lipid (the mass ratio of protein to lipid is 7:3), the corrected molecular weight of the isolated unit of highest purity comes to 1.4 X 106--in fair agreement with the value of 1.3 X 106 computed for the molecular weight of the elementary particle in situ. It is possible to calculate the theoretical molecular weight of the electron transfer chain from the known molecular weights of each of the component oxidation-reduction proteins and from the molecular proportions of these proteins. Such a calculation is summarized in Table V. Implicit in this calculation is the assumption that the electron transfer chain does not contain any structural protein. The computed molecular weight comes to a value of 0.9 X 106 on a protein basis and corresponds closely with that for the elementary it can be examined in the ultracentrifuge, and the asymmetry and shape of this derived unit are not necessarily indicative of the shape either of the isolated unit before "solubilization" or of the unit in situ before extraction from the mitochondrial membranes. The fact that the elementary particle is at once external to, and part of, the outer membrane of the crista probably explains why simple sonication does not provide an effective means for separating the elementary particle from the structural protein-lipid network.
Identification
of
the
Isolated
Elementary
Particle with That Seen i n Situ

Evidence that the isolated elementary particle is of the appropriate dimensions to correspond with the particle seen in situ merely eliminates the problem of size as a barrier to the identification, but is not necessarily direct proof for the identifi-
94
TaE JOURNAL OF CELL BIOLOGY VOLUME ~$, 1964
cation. We have, in fact, no direct proof for this identification if direct proof is possible; however, there is a wealth of circumstantial evidence which collectively makes the identification highly probable. The unit carrying the electron transfer chain can be purified some fivefold relative to the mitochondrion. Therefore, any structure presumed to be associated with electron transfer activity should not account for more than 20 per cent of the total dry weight of the mitochondrion. The particle visualized in situ would certainly fulfill this requirement. The isolated electron transfer unit, as well as the purified complexes, contains characteristic electron opaque substructures. Under favorable conditions similar substructures can also be seen in the head piece of the particle in situ. The mitochondrion contains a variety of particles that are concerned in implementing citric cycle oxidations, fatty acid oxidation, fatty acid synthesis, etc. When mitochondria are irradiated with sonic oscillations, the particles concerned with these auxiliary mitochondrial functions can, in large measure, be separated from the particles implicated in the electron transfer function. After such a separation, the presence of the characteristic elementary particles in the fraction retaining electron transfer activity is demonstrable. At least in heart muscle there is no evidence that integrated electron transfer activity can be observed in fractions (prepared by sonic irradiation) that are devoid of the characteristic particles. When bile salts are used in the isolation of the particles with electron transfer activity, there is a major change in the size and shape of the resulting particles. The characteristic appearance of the elementary particle can then no longer be used as a guide for identification. The more active a mitochondrion (per mg of protein), the more numerous are the cristae per mitochondrion and the greater is the number of the particles in question. Thus, there is a close correspondence between electron transfer activity and the number of native elementary particles. All lines of available evidence point to the identification of the isolated particles with those seen in situ, but in absence of direct evidence the identification must remain provisional.
Interpretation of the Arrangement of Parts in the Native Particle

The electron transfer chain is made up of four interdigitating enzymic complexes which can re-
assemble spontaneously to form the original electron transfer chain in 1 : 1 : 1 : 1 molecular proportions (44, 45, 47). How can this interrelationship of the four segments of the chain be rationalized in terms of the structural arrangement of the parts of the native elementary particle? The head piece could be assigned as the locale of complex IV (the cytochrome oxidase); the stalk as the locale of complex I I I (the QH~ cytochrome c reductase); and the base piece as the locale of complexes I (DPNH-coenzyme Q. reductase) and II (succinic-coenzyme Q reductase). This tentative assignment of the four complexes is based on biochemical considerations. Complexes I and II must interact with DPNH and succinate, respectively, both of which are localized in the interior of the crista, whereas complex IV must interact with molecular oxygen which would be more readily available in the solution outside the crista rather than in its interior. The cytochrome oxidase appears to be the largest of the four complexes (cf. Table VI). The molecular weight of the cytochrome oxidase complex of highest purity is 420,000 on a protein basis. Normalized for 30 per cent lipid, that would come to a molecular weight of 600,000. A sphere of density 1.25 with a molecular weight of 600,000 would have a diameter of about 115 A. The dimensions of the head piece of the particle in situ (108 A after 12 per cent correction for underestimation) would not be incompatible with the assignment of the cytochrome oxidase to the head piece. Electron micrographs of isolated and purified cytochrome oxidase show an annular repeating unit about 80 A in diameter made up of structured subunits in the periphery and electron opaque material in the core. More remains to be done, but a case can be made for the provisional identification of cytochrome oxidase with the head piece of the native particle. Perhaps the best basis for comparison would be the experimentally determined diameters for the head piece (80 to 100 A) and the corresponding range for cytochrome oxidase (70 to 90 A). Cytochrome oxidase is a polymeric complex. The exact dimensions of the polymer unit may be affected by a variety of factors. Complexes I, II, and III have molecular weight of 300,000 or less. The electron microscopic examination of these particular complexes has not been carried very far. None of them shows the marked regularity of the repeating units of the cytochrome oxidase. The available data suggest
FERN/~NDEZ-MOR~xNET AL. Macromolecular Repeating Unit of Mitochondrial Structure
95
that the repeating units of complexes I, II, and III are considerably smaller than those of the cytochrome oxidase (diameters in the 40 to 60 A range), The measurements of size, tentative though they may be, are not incompatible with the assignment of complexes I and II to the base piece and of complex I I I to the stalk.
Demonstration of Elementary Particles in other Laboratories

Since our preliminary announcement in 1961 of the presence of repeating particles of 80 to 100 A in diameter in the mitochondrial membranes (25, 34), there has been confirmation from several laboratories of our major findings. Parsons (55, 56) has done particularly fine work in devising new methods for negative staining of mitochondria and tissue preparations with phosphotungstate. The uniformity of the elementary particles in his mitochondrial preparations, and the clear visualization of the stalks, are indeed remarkable. Parsons has also concerned himself with the particles of the external envelope. He has postulated a new kind of unit peculiar to the external layer of the envelope (56). The elegant studies of Smith (70, 71) on the sarcosomes of insect flight muscle are of particular importance because these specialized mitochondria can be isolated from the fresh muscle fibers with a minimum of preparative manipulation. The presence of particles in the sarcosomes of the flight muscle (after staining with phosphotungstate) which were similar to those we have described was fully confirmed by Smith. Stoeckenius (72) has also presented extensive electron microscopic evidence of the repeating particles and their stalks in mitochondrial preparations.
Collateral Evidence of Subunit Organization of M itochondrial Membranes

Suggestions of regular repeating structures in the planes of unit membranes after various preparatory techniques have been obtained by others. Pease (57) has recently found a very regular pattern of dense beads repeating at a period of about 160 A along two closely apposed outer mitochondrial membranes in a spherule of cat retina. Very recently Robertson (62) has demonstrated by electron microscopy that the synaptic discs in transverse section show a central beading repeating at a period of about 85 A associated with
scalloping of the cytoplasmic surfaces. In frontal views, an hexagonal array of close-packed polygonal facets is seen. These repeat in a period of about 95 A. Each has a central dense spot about 25 A in diameter. Similar subunits are seen in the unit membranes of synaptic vesicles. Suggestions of a similar hexagonal pattern have also been found in mitochondrial membranes by Robertson (62). Whether or not his "geodesic" pattern represents a macromolecular pattern in intact native membranes remains to be determined. Further interpretations must await evidence from other technical approaches, notably x-ray diffraction studies of membranes. Important collateral morphological evidence in support of the existence of a subunit structure of the membranes is afforded by the comprehensive studies of the transformations of mitochondria during spermatogenesis (2). As Andr6 (2) has clearly demonstrated in extensive electron microscopic studies in nearly 50 widely varied species, mitochondria exhibiting normal ultrastructure at the beginning of spermatogenesis undergo progressive modifications to derivative structures of unusual pattern. These transformations are particularly striking in Testacella. In this gastropod, a true metamorphosis occurs during which membranes, matrix, pseudomatrix, and dense granules disappear, while orderly domains appear. These orderly domains expand and finally form two concentric muffs of paracrystalline structure. The constituent elements of this paracrystalline network are "hollow rodlets measuring 90 A in diameter." Andr6 suggests that these paracrystalline structures "represent a particular state of the respiratory proteins" (2) and points out that the dimensions of the units of this paracrystalline structure approach the dimensions of the EP's described in our studies. Similar modifications or transformations of mitoehondrial membranes in the course of physiological changes in various types of mitochondria have also been reported by Favard and Carasso (15), Fern~ndez-Morfin (21, 22), and others.
The Oxysome Concept

Chance, Estabrook, and Lee (10) have recently interpreted the electron micrograph finds of H. Fernfindez-Morhn and Parsons and the biochemical findings of our group in terms of the "oxysome" hypothesis. They consider that the particle which we have isolated may be an artifact
96
THE JOURNAL OF CELL BIOLOGY VOL~IME ~!~, 1964
arising from the fusion of the individual oxidationreduction protein components of the electron transfer chain with each other as with the primary dehydrogenases and the coupling factors. According to Chance et al. (10), the native particles correspond to the component proteins of the electron transfer chain and to those of associated processes. Thus, one such particle might be cytochrome b, one cytochrome cl, etc. The smallest molecular unit of the electron transfer chain contains one molecule each of cytochrome el, succinic dehydrogenase, and DPNH dehydrogenase, three molecules of cytochrome b, six molecules of cytochrome a, and some 18 molecules of non-heine iron protein. At least 15 more proteins would have to be involved to include the primary dehydrogenase systems and the coupling factors. One oxysome would have to encompass over 50 separate particles. There are three major objections to this interpretation. First, there is not a single protein component of the electron transfer chain large enough to account for the size of the particle visualized in situ. A molecule of cytochrome b has a molecular weight of 28,000 according to Goldberger, Bomstein, and Tisdale (31). For an assumed spherical particle, the maximal diameter of cytochrome b would be about 40 A. The corresponding value for cytochrome Cl (molecular weight 38,000 according to Bomstein et al., 6) would be about 45 A; and for the nonheme iron protein, 40 A (molecular weight 25,000 according to J. Rieske et al., 60). Second, the oxysome hypothesis involves the concept of molecular collisions between individual protein molecules of the chain--a concept that is incompatible with a vast amount of biochemical knowledge, in particular that concerning the arrangement of the complexes of the electron transfer chain, each a mosaic or fusion of four or more proteins. Within the solid state matrix of these complexes, there is no evidence of anything resembling molecular collision between adjacent proteins. Third, the dimensions of the individual proteins of the chain are too far removed from the dimensions of the particles seen in situ to justify serious consideration of the individual protein hypothesis.
the coupling function that are incorporated into the fabric of the mitochondrial membrane. Like Chance, Lehninger conceives of the electron transfer chain as a linear array of cytochromes and other oxidation-reduction proteins interacting one with the other by molecular collision. His estimate of the dimensions of the respiratory assembly comes to a molecular weight of 1.5 million, and he assumes that the respiratory chain accounts for 20 per cent of the total membrane (50, 51). Lehninger has never intended his respiratory assembly hypothesis to do more than serve as a model, since it is not based on direct experimental evidence. But implicit in the hypothesis is the notion that the respiratory assembly would not be separable from the mitochondrial membrane. The isolation of the elementary particle some fivefold more concentrated in oxidation-reduction components than the original mitochondrion establishes this separability.
Universality of the Concept of the Elementary Particle

Elsewhere we have developed the thesis that all membrane systems of living cells contain particles analogous to the elementary particles of the mitochondrion (18, 25, 35). Sauer and Calvin (63) have isolated the quantasome from the chloroplast, which is similar in dimensions to the elementary particle of the mitochondrion. Hultin et al. (48) have demonstrated the intimate association of the glycolytic complex of enzymes with the membrane of the red blood cell. The particles which carry these activities may correspond to the particles seen in situ in the membrane of the red blood corpuscles when stained with phosphotungstate (Fernhndez-Morfin, unpublished studies). The microsomal membrane contains a particulate unit having a complete electron transfer chain (DPNH flavoprotein, cytochrome b, high lipid content, etc.). These developments suggest that the mitochondrion is not unique; particles and complexes with comparable functions are present in other membranes, and thus the elementary particle (considered in its broadest sense) may be one of the universals of all membrane systems. Regardless of alternative possibilities, the major conclusion to be derived from these correlative studies is that a repeating unit of mitochondrial structure, first disclosed by electron microscopy, has led to the isolation of a compact multi-enzyme
The Respiratory Assembly of Lehninger

Lehninger (50) has for some years visualized the respiratory assembly as a set of oxidationreduction proteins and proteins concerned with
FERN~NDEz-MoI~.~.N ET AL, MacromolecularRepeating Unit of Mitochondrial Structure
97
entity which is similar in many respects to the native subunit. The combined approach, linking ultrastructure with biochemistry at the same level of resolution, for the first time invested a lamellar system (the mitochondrial membrane) with specific enzymic detail. W h a t remains to be done may be summarized under five headings: (1) the direct identification of the isolated and native elementary particles; (2) the identification of the three parts of the native elementary particle with the four complexes of the electron transfer chain; (3) the delineation of the gross molecular structure of the individual complexes; (4) the characterization of the particles associated with the outer membrane of the external envelope as functionally and structurally different from the elementary particles associated with the inner membrane and the cristae; (5) the isolation and characterization of elementary particles from other membrane systems. These are all problems that could not even be formulated before the discovery of the elementary particle. The combination of the biochemical and electron microscopic approaches has been decisive in opening the door to the molecular domain where structure and function merge. We would like to think that the first few lines of an exciting new chapter have been contributed in the present communication. We are greatly indebted to Professors John Anderegg, William Beeman, and Robert Book for their many REFERENCES 1. AMBE, K., and VENKETARAMAN, A., Biochem. and Biophys. Research Commun., 1959, 1, 133. 2. ANDRf~, J., J. Ultrastruct. Resemch, suppl., 1962, 3, 1. 3. BANOHAM, A. D., and HORNE, R. W., Nature, 1962, 196, 952. 4. BARRNeTT,R. J., and PALADE,G. E., J. Biophys#. and Biochem. Cytol., 1957, 3,577. 5. BLAIR, P. V., ODA, T., GREEN, D. E., and FERN~NDEZ-MoR~,~, H., Biochemistry, 1963, 2, 766. 6. BOMSTEIN, R., GOLDBERGER, R., and TISDALE, H., Biochim. et Biophysica Acta, 1960, 50, 527. 7" BRENNER, S., and HORNE, R. W., Biochim. et Biophysica Acta, 1959, 34, 103. 8. CASPAR, D. L. D., and KLUG, A., Cold Spring Harbor Symp. Quant. Biol., 1962, 27, 1. 9. CHAZqC~, B., and WILLIAMS, G. R., Advances Enzymol., 1956, 17, 65.
suggestions, to Dr. Pauline Yang for carrying out the ultracentrifuge runs, to Donald Silver for his technical assistance, and to Oscar Meyer & Company of Madison, Wisconsin, for kindly providing fresh beef hearts. Sincere thanks are also due to Frederick B. Merk and Charles Hough for expert technical assistance, and to Sandra J. Riddle and Janice Foss for their help in preparing the manuscript. The authors gratefully acknowledge stimulating discussions with Professors Francis O. Schmitt, Neurosciences Research Program, George Palade of The Rockefeller Institute, William H. Sweet of Massachusetts General Hospital, and William Bloom of the University of Chicago. The substance of this communication was reported in joint papers at the Annual Meeting, National Academy of Sciences, April, 1963, and the abstracts of which have been published in Science, 1963, 140, 381 and 382. The work at the Institute for Enzyme Research was supported in part by the Division of General Medical Sciences, graduate training grant 2G-88; the National Heart Institute, research grant HE00458, United States Public Health Service; the National Science Foundation, grant G-3227; and the Atomic Energy Commission, contract No. AT (ll-l)-l151. The work at the Mixter Laboratories mad The Department of Biophysics was supported by United States Atomic Energy Commission contract AT(301)-2278; by grants B-2460, C-3174, and NB-04267 from the National Institutes of Health. Received for publication, November 1, 1963.
10. CHANCa~,B., ESTABROOK,R. W., and LEE, C. P . Science, 1963, 194, 379. 11. CRANE, F. L., GLENN, J. L., and GREEN, D. E., Biochim. et Biophysica Acta, 1956, 22, 475. 12. CRIBDLE, R. S., Bocx, R. M., GREEN, D. E., and TISDA~, H., Biochem. and Biophys. Research Commun., 1961, 5, 75. 13. CRmD~, R. S., BoeR, R. M., GREEN, D. E., and TISDAI~, H., Biochemistry, 1962, 1, 827. 14. DASS, C., ILLs, H., ZIEO~R, D. M., LINNANE, A. W., and GREEN, D. E., Biochim. et Biophysica Acta, 1958, 28, 524. 15. FAVARD, P., and CARaSSO, N., Arch. Anat. mitt. et morphol. Exp., 1958, 47, 221. 16. FERN~NDEz-MoR~N, H., Ark. Zool. K. Svenska Vetenskapsakademien, 1948, 40A, 1. 17. FERNI,NDEz-MoRN, H., Exp. Cell Research, 1952, 3, 282.
98
THE JOURNAL OF CELL BIOLOGY VOLUME ~ , 1964
18. FERNANDEZ-MORAN,H., Rev. Modern Phys., 1959, 31, 319. 19. FERNANDEZ-MORAN, H., Ann. New York Acad. Sc., 1960, 85, 689. 20. FERNANDEZ-MORAN, H., J. Appl. Phys., 1960, 31, 1840. 21. FERNANDEZ-MORN, H., in Macromolecular Complexes, (M. V. Edds, editor), New York, The Ronald Press Co., 1961, 113. 22. FERNANDEZ-MoRAN,H., in The Structure of the Eye, (G. K. Smelser, editor), New York, Academic Press, Inc., 1961, 521. 23. FERNANDEZ-MoRAN, H., Research Publ. Assn. Nero. and Ment. Dis., 1962, 40, 235. 24. F~RNkNDEZ-MORAN,H., in The Interpretation of Ultrastructure, (R. Harris, editor), Academic Press, Inc., New York and London, 1962, 411. 25. FERNkNDEz-MoRXN, H., Circulation, 1962, 26, 1039. 26. FERNANDEZ-MoRAN, H., in Macromolecular Specificity and Biological Memory, (F. O. Schmitt, editor), Cambridge, Massachusetts Institute of Technology Press, 1962, 39. 27. FERNkND~Z-MoRAN,H., Science, 1963, I40, 381. 28. FERNANDEZ-MORkN,H., in Tercentenary of the Microscope in Living Biology, Royal Microscopical Society Journal, 1964, 83, 183. 29. FERNANDEZ-MORAN,H., and FINEAN,J. B., 3.. Biophysic. and Biochem. Cytol., 1957, 3, 725. 29a. FERNANDEZ-MORAN, H., and SCHRAMM, G., Z. Naturforsvh., 1958, 13b, 68. 29b. FLEISCHER, S., KLOUWEN, H., and BRIERLEY, G., 3". Biol. Chem., 1961, 236, 2936. 30. FOWLER, L., and RICHARDSON, S. J., J. Biol. Chem., 1963, 238, 456. 31. GOLDBERGER, R., BOMSTEIN, R., and TISDALE, H., J. Biol. Chem., 1961, 236, 2788. 32. GREEN, D. E., in Harvey Lectures (Series 52, 1956-57), New York, Academic Press, Inc., 1958, 177. 33. GREEN, D. E., Radiation Research, suppl., 1960, 2, 504. 34. GREEN, D. E., Plenary Lecture, Sixth International Congress of Biochemistry, Moscow, 1961, Preprint No. 176. 35. GREEN,D. E. Scient. American, 1964, 210, 63. 36. GREEN, D. E., LOOMIS, W. F., and AtmREACH, V. H., J. Biol. Chem., 1948, 172, 389. 37. GREEN, D. E., and I-tATF-FI,Y., Science, 1961, 133, 13. 38. GREEN,D. E., and ODA, T. J., J. Biochem., 1961, 49, 743. 39. GREEN, D. E., TISDALE, H., CRIDDLE, R. S., and BooK, R. M., Biochem. and Biophys. Research Commun., 1961, 5, 81. 40. GREEN,D. E., and FLEISCX-mR,S., in Horizons in
41. 42. 43. 44.
45.
46. 47.
48.
49. 50. 51. 52. 53. 54. 55.
56. 57. 58.
59. 60. 61. 62. 63. 64. 65. 66.
Biochemistry, (M. Kasha and B. Pullman, editors), New York and London, Academic Press, Inc., 1962, 381. GREEN, D. E., and WHARTON,D., Biochem. Z., 338, 1963. GRIFFITHS,D. E., and WHARTON,D., J. Biol. Chem., 1961, 236, 1850. HAGmARA, B., 12th Symposium on Enzyme Chemistry, Japan, 1960, 140. HATEFI, Y., HAAVm, A. G., and GRIFFITnS, D. E., Biochem. and Biophys. Research Commun., 1961, 4, 441,447. HATEFI, Y., HAAVIK, A. G., FOWLER, L. R., and GRIVFITHS,D. E., J. Biol. Chem., 1962, 237, 2661. HATEFr, Y., HAAVIK, A. G., and GRIFFITHS, D. E., J. Biol. Chem., 1962, 237, 1676. HATEFI,Y., HAAVIK, A. G., JURTSHUK, P., and GRIFFITHS, D. E., or. Biol. Chem., 1962, 237, 1676. HULTIN, H., RICHARDSON, S., SALMON, B., MURER, E., BRmRLEV, G., and GREEN, D. E., in preparation. LEHNINGER, A. L., and KENNEDY, E. P., J. Biol. Chem., 1949, 179, 957. LEHNINOER, A. L., Rev. Modern Phys., 1959, 31, 136. LEHmNGER, A. L., Physiol. Rev., 1962, 42, 467. MEROLA,J., and COLEMAN, R., in preparation. PALADE, G. E., Anat. Rev., 1952, 114, 427. PALADE, G. E., J. Histoehem. and Cytochem., 1953, 1, 188. PARSONS, D. F., Proceedings 5th International Congress Electron Microscopy, (S. S. Breese, editor), New York, Academic Press, Inc., 1962, 2, abstract X-1. PARSONS, D. F., Science, 1963, 140, 985. PEASE, D. C., J. Cell Biol., 1962, 15, 385. 1LEEO, L. S., Chemistry and function of lipoic acid. in Comprehensive Biochemistry, 1963, Elsevier Publ. Co., Amsterdam, 3. REVEL, J. P., FAWC~TT,D. W., and PHILPOTT, C. W., J. Cell Biol., 1963, 16, 187. RIESKE,J., McLaNNAN, D., COLEMAN, R., and YANO, P., in preparation. tlaESX~, J., ZAUOO, W., and WHARTON, D., in preparation. ROBERTSON,J. D., J. Cell Biol., 1963, 19, 201. SALTER, K., and CALVIN, M., J. Mol. Biol., 1962, 4, 451. SCHMITT, F. O., Develop. Biol., 1963, 7, 546. SILK, rITZ, P., and WATSON,M. L., J. Biophysic. and Biochem. Cytol., 1956, 2, 653. SINGER, T. P., KEARNEY, E. D., and MASSEY, V., in Enzymes: Units of Biological Structure and Function, (O. H. Gaebler, editor),
FERNANDEz-MoRAN ET AL. Macromolecular Repeating Unit of Mitochondrial Structure
99
67. 68. 69. 70. 71.
72.
73. 74. 75.
New York, Academic Press. Inc., 1956, 185, 215. SJt~STRAND,F. S., Nature, 1953, 171, 303. SJ6STRAND, F. S., Radiation Research, suppl., 1960, 2, 349. SMITH,A., and HANSEN,M., Biochim. et Biophysica Acta, 1963. SMITH,D. S., 3". Cell Biol., 1963, 19, 115. SMITH,D. S., personal communication. STOECKENIIJ$,W., J. Cell Biol., 1963, 16, 483. TAra~MORI, S., SEKUZU, I., and OKVNUKI, K., Nature, 1960, 188, 593. THEORELL, H., and AKESON, A., J. Am. Chem. Soc., 1941, 63, 1804. TISDALE,H., TZAGALOFF,A., and GREEN, D. E., in preparation.
75a. TISDALE,H., WHARTON,D., and GREEN,D. E., Arch. Biochem. and Biophys., 1963. 76. WATSON, M. L., and SIEKEVITZ, P., J. Biophysic. and Biochem. Cytol., 1956, 2, 639. 77. WHARTON, D., and TZAGOLOFF, A., Biochem. and Biophys. Research Commun., 1963, 13, 121. 78. WOLKEN,J. J., J. opt. Soc. America, 1963, 53, 1. 79. ZIEOLER, D. M., LINNANE,A. W., GREEN, D. E., DASS, C. M. S., and RIs, H., Biochim. et Biophysica Acta, 1958, 28, 524. 80. ZIEOLER, D. M., and LINNANE, A. W. Biochim. et Biophysica Acta, 1958, 30, 53. 81. ZXEOLER,D. M., and DOEG, K. A., Arch. Biochem. and Biophys., 1962, 97, 41.
100
THE JOURNAL OF CELL BIOLOGY VOLUME ~ ,
1964

Articulo HFM

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Articulo HFM

Uploaded by

Copyright:

Available Formats

Published July 1, 1964

Downloaded from jcb.rupress.org on February 4, 2011

Published July 1, 1964

Downloaded from jcb.rupress.org on February 4, 2011

THE JOURNALOF CELL BIOLOGY" VOLUME~ , 1964

Published July 1, 1964

Downloaded from jcb.rupress.org on February 4, 2011

FERN~.NDEz-MoR~.N ET AL. Macromolecular Repeating Unit of Mitochondrial Structure

Published July 1, 1964

Downloaded from jcb.rupress.org on February 4, 2011

THE JOURNAL OF CELL BIOLOOY - VOLUME ~ ,

Published July 1, 1964

Method/or Large Scale Preparation of Mitochondria in 0.5 M Sucrose

Downloaded from jcb.rupress.org on February 4, 2011

FERN.~NDEZ-MOR,~N ST AL. Macromolecular Repeating Unit of Mitochondrial Structure

Published July 1, 1964

Negative Staining Techniques

Downloaded from jcb.rupress.org on February 4, 2011

Index to Electron Micrographs

THB JOURNAL OF CELL BIOLOGY VOLUME 22, 1964

Published July 1, 1964

Downloaded from jcb.rupress.org on February 4, 2011

Macromolecular Repeating Unit of Mitochondrial Structure

Published July 1, 1964

Downloaded from jcb.rupress.org on February 4, 2011

Tn~ JOURNAL OF CELL BIOLOGY VOLUME ~ , 1964

Published July 1, 1964

Downloaded from jcb.rupress.org on February 4, 2011

MacromolecularRepeating Unit of Mitochondrial Structure

Published July 1, 1964

2. Thin Sectioning Techniques

Downloaded from jcb.rupress.org on February 4, 2011

THE JO~I~NAL Or" CELL BIOLOGY VOLU~E 2~, 1964

Published July 1, 1964

Downloaded from jcb.rupress.org on February 4, 2011

Macromolecular Repeating Unit of Mitochondrial Structure

Published July 1, 1964

Electron Microscopy of Isolated Mitochondria

Downloaded from jcb.rupress.org on February 4, 2011

THE JOURNAL OF CELL BIOLOGY VOLUME g~, 1964

Published July 1, 1964

Downloaded from jcb.rupress.org on February 4, 2011

Macromolecular Repeating Unit of Mitoctwndrial Structure

Published July 1, 1964

Downloaded from jcb.rupress.org on February 4, 2011

THE JOm~AL OF CELL BIOLOGY VOLUM~ ~ , 1964

Published July 1, 1964

Downloaded from jcb.rupress.org on February 4, 2011

Macromolecular Repeating Unit of Mitochondrial Structure

Published July 1, 1964

Downloaded from jcb.rupress.org on February 4, 2011

THE JOURNAL OF CELL BIOLOGY VOLUME 22, ]964

Published July 1, 1964

Downloaded from jcb.rupress.org on February 4, 2011

MacromolecularRepeating Unit of Mitochondrial Structure

Published July 1, 1964

Downloaded from jcb.rupress.org on February 4, 2011

TIlE JOUENAL OF CELL BIOLOGY VOLUME ~ ,

Published July 1, 1964

Downloaded from jcb.rupress.org on February 4, 2011

MacromolecularRepeating Unit of Mitochondrial Structure

Published July 1, 1964

DPNH-coenzyme Q reductase* Succinic-coenzyme Q reduet ase ~ QH~-cytoehrome c reduetase Cytoehrome oxidase][